When using AlphaFold Colab (https://colab.research.google.com/github/deepmind/alphafold/blob/main/notebooks/AlphaFold.ipynb), I am getting the following error on step 4 ("Search against genetic databases"):
KeyError Traceback (most recent call last)
<ipython-input-12-a4ead9510dc9> in <module>
145 for db_name, db_results in raw_msa_results.items():
146 merged_msa = notebook_utils.merge_chunked_msa(
--> 147 results=db_results, max_hits=MAX_HITS.get(db_name))
148 if merged_msa.sequences and db_name != 'uniprot':
149 single_chain_msas.append(merged_msa)
1 frames
/opt/conda/lib/python3.7/site-packages/alphafold/notebooks/notebook_utils.py in merge_chunked_msa(results, max_hits)
105 e_values_dict = parsers.parse_e_values_from_tblout(chunk['tbl'])
106 # Jackhmmer lists sequences as <sequence name>/<residue from>-<residue to>.
--> 107 e_values = [e_values_dict[t.partition('/')[0]] for t in msa.descriptions]
108 chunk_results = zip(
109 msa.sequences, msa.deletion_matrix, msa.descriptions, e_values)
/opt/conda/lib/python3.7/site-packages/alphafold/notebooks/notebook_utils.py in <listcomp>(.0)
105 e_values_dict = parsers.parse_e_values_from_tblout(chunk['tbl'])
106 # Jackhmmer lists sequences as <sequence name>/<residue from>-<residue to>.
--> 107 e_values = [e_values_dict[t.partition('/')[0]] for t in msa.descriptions]
108 chunk_results = zip(
109 msa.sequences, msa.deletion_matrix, msa.descriptions, e_values)
KeyError: '0000|query')
Notably, this error seems to be dependent on the input sequence, because I get it with some sequences but not with others. I am using ColabPro and I have tried to tweak parameters like the amount of computing power but it did not help. I have never experienced this error before switching to ColabPro. Any help would be appreaciated.